Matches in SemOpenAlex for { <https://semopenalex.org/work/W3183432322> ?p ?o ?g. }
- W3183432322 abstract "Demonstration-guided reinforcement learning (RL) is a promising approach for learning complex behaviors by leveraging both reward feedback and a set of target task demonstrations. Prior approaches for demonstration-guided RL treat every new task as an independent learning problem and attempt to follow the provided demonstrations step-by-step, akin to a human trying to imitate a completely unseen behavior by following the demonstrator's exact muscle movements. Naturally, such learning will be slow, but often new behaviors are not completely unseen: they share subtasks with behaviors we have previously learned. In this work, we aim to exploit this shared subtask structure to increase the efficiency of demonstration-guided RL. We first learn a set of reusable skills from large offline datasets of prior experience collected across many tasks. We then propose Skill-based Learning with Demonstrations (SkiLD), an algorithm for demonstration-guided RL that efficiently leverages the provided demonstrations by following the demonstrated skills instead of the primitive actions, resulting in substantial performance improvements over prior demonstration-guided RL approaches. We validate the effectiveness of our approach on long-horizon maze navigation and complex robot manipulation tasks." @default.
- W3183432322 created "2021-08-02" @default.
- W3183432322 creator A5028858301 @default.
- W3183432322 creator A5035555714 @default.
- W3183432322 creator A5048121389 @default.
- W3183432322 creator A5054155451 @default.
- W3183432322 date "2021-07-21" @default.
- W3183432322 modified "2023-09-27" @default.
- W3183432322 title "Demonstration-Guided Reinforcement Learning with Learned Skills" @default.
- W3183432322 cites W1986014385 @default.
- W3183432322 cites W1999874108 @default.
- W3183432322 cites W2005789722 @default.
- W3183432322 cites W2031261758 @default.
- W3183432322 cites W2164233973 @default.
- W3183432322 cites W2167224731 @default.
- W3183432322 cites W2566467060 @default.
- W3183432322 cites W2729615412 @default.
- W3183432322 cites W2736601468 @default.
- W3183432322 cites W2741122588 @default.
- W3183432322 cites W2753738274 @default.
- W3183432322 cites W2785342287 @default.
- W3183432322 cites W2890026535 @default.
- W3183432322 cites W2902592652 @default.
- W3183432322 cites W2902711054 @default.
- W3183432322 cites W2904246096 @default.
- W3183432322 cites W2920215304 @default.
- W3183432322 cites W2962902376 @default.
- W3183432322 cites W2962936820 @default.
- W3183432322 cites W2962957031 @default.
- W3183432322 cites W2963099939 @default.
- W3183432322 cites W2963277051 @default.
- W3183432322 cites W2963376229 @default.
- W3183432322 cites W2963411833 @default.
- W3183432322 cites W2963508354 @default.
- W3183432322 cites W2963713397 @default.
- W3183432322 cites W2964121744 @default.
- W3183432322 cites W2994689640 @default.
- W3183432322 cites W2995107092 @default.
- W3183432322 cites W2995706821 @default.
- W3183432322 cites W2996668373 @default.
- W3183432322 cites W2996698477 @default.
- W3183432322 cites W3016525976 @default.
- W3183432322 cites W3027456239 @default.
- W3183432322 cites W3028830971 @default.
- W3183432322 cites W3031840745 @default.
- W3183432322 cites W3034786558 @default.
- W3183432322 cites W3048659003 @default.
- W3183432322 cites W3096964654 @default.
- W3183432322 cites W3097907450 @default.
- W3183432322 cites W3107295910 @default.
- W3183432322 cites W3122543654 @default.
- W3183432322 cites W3210184670 @default.
- W3183432322 hasPublicationYear "2021" @default.
- W3183432322 type Work @default.
- W3183432322 sameAs 3183432322 @default.
- W3183432322 citedByCount "0" @default.
- W3183432322 crossrefType "posted-content" @default.
- W3183432322 hasAuthorship W3183432322A5028858301 @default.
- W3183432322 hasAuthorship W3183432322A5035555714 @default.
- W3183432322 hasAuthorship W3183432322A5048121389 @default.
- W3183432322 hasAuthorship W3183432322A5054155451 @default.
- W3183432322 hasConcept C107457646 @default.
- W3183432322 hasConcept C119857082 @default.
- W3183432322 hasConcept C127413603 @default.
- W3183432322 hasConcept C154945302 @default.
- W3183432322 hasConcept C165696696 @default.
- W3183432322 hasConcept C177264268 @default.
- W3183432322 hasConcept C199360897 @default.
- W3183432322 hasConcept C201995342 @default.
- W3183432322 hasConcept C2780451532 @default.
- W3183432322 hasConcept C38652104 @default.
- W3183432322 hasConcept C41008148 @default.
- W3183432322 hasConcept C90509273 @default.
- W3183432322 hasConcept C97541855 @default.
- W3183432322 hasConceptScore W3183432322C107457646 @default.
- W3183432322 hasConceptScore W3183432322C119857082 @default.
- W3183432322 hasConceptScore W3183432322C127413603 @default.
- W3183432322 hasConceptScore W3183432322C154945302 @default.
- W3183432322 hasConceptScore W3183432322C165696696 @default.
- W3183432322 hasConceptScore W3183432322C177264268 @default.
- W3183432322 hasConceptScore W3183432322C199360897 @default.
- W3183432322 hasConceptScore W3183432322C201995342 @default.
- W3183432322 hasConceptScore W3183432322C2780451532 @default.
- W3183432322 hasConceptScore W3183432322C38652104 @default.
- W3183432322 hasConceptScore W3183432322C41008148 @default.
- W3183432322 hasConceptScore W3183432322C90509273 @default.
- W3183432322 hasConceptScore W3183432322C97541855 @default.
- W3183432322 hasLocation W31834323221 @default.
- W3183432322 hasOpenAccess W3183432322 @default.
- W3183432322 hasPrimaryLocation W31834323221 @default.
- W3183432322 hasRelatedWork W1563109146 @default.
- W3183432322 hasRelatedWork W2580475959 @default.
- W3183432322 hasRelatedWork W2767506186 @default.
- W3183432322 hasRelatedWork W2793939502 @default.
- W3183432322 hasRelatedWork W2889681745 @default.
- W3183432322 hasRelatedWork W2963099939 @default.
- W3183432322 hasRelatedWork W2963129648 @default.
- W3183432322 hasRelatedWork W2963176272 @default.
- W3183432322 hasRelatedWork W2963619650 @default.
- W3183432322 hasRelatedWork W2981701630 @default.